Hub Options

Control agent selection, voice settings, and conversation management

Hub Options

Access comprehensive control over your avatar interactions through the options menu. Configure agents, select voices, choose AI models, and manage conversation history from a single interface.

Accessing Options

Options Menu

Click the gear icon in the top-left corner of the Hub (next to Chat and Talk buttons) to open the options panel. This menu is available in both chat and voice modes.

Options Panel Components

The options menu provides four main control sections:

• Agent Selection
• Voice Settings
• LLM Model Selection
• Conversation History

Options panel showing agent dropdown, voice selection, and conversation history

Options Panel Interface Variations

Complete Options Menu:

Full options panel showing all available controls: agent selection, voice settings, LLM models, and conversation history

Chat Response Interface:

Options panel integrated with chat response showing agent selection during active conversation

Video Scene Upload Options

When using the Video Wall scene, access video upload functionality through the options controls:

Video upload dialog showing drag-and-drop interface for scene video customization with MP4 format requirements

Video Upload Features

Upload Specifications:

• File Format: MP4 videos only
• Size Limit: Maximum 50MB per video
• Upload Method: Drag and drop interface or file browser selection
• Scene Integration: Videos play as background in Video Wall scene

Upload Process:

1. Access Video Wall Scene: Select Video Wall scene from scene menu
2. Open Upload Dialog: Click upload button near Chat/Talk controls
3. Select Video File: Drag MP4 file or browse to select
4. Upload Confirmation: Wait for upload completion confirmation
5. Automatic Playback: Video begins playing as scene background

Agent Selection

Agent Dropdown

Available Agents:

• All Agents Available: Worker Agents and Presenter Agents appear in the same dropdown.
• Current Selection: The active agent is highlighted so you know which experience is live.
• Agent Types: Worker Agents cover full knowledge, realtime, and integrations; presenter agents focus on a single deck.
• Real-Time Switching: Change agents instantly during conversations without refreshing the session.

Agent dropdown showing Worker and presenter agents like Demo Worker, Support Worker, and Product Presenter

Agent Switching Benefits

Specialized Experiences:

• Different Expertise: Worker Agents can focus on support, sales, or internal knowledge while presenter agents guide demos.
• Contextual Switching: Match the agent to the current topic or audience on the fly.
• Seamless Transition: Maintain conversation context even when swapping agents mid-session.
• Instant Access: Switch as often as needed; changes apply immediately.

Agent-Specific Features

Knowledge Sources:

• Custom Training: Worker Agents draw from the documents, URLs, and uploads you assign.
• Presentation Mode: Presenter agents rely on a single PDF to keep the story aligned.
• Full Toolkit: Both agent types can search the web, call integrations, and generate media in realtime.
• Unique Responses: Each agent has its own tone, behavior settings, and available tools.
• Integration Access: Configure integrations per agent to tailor available actions.

Voice Selection

Voice Families by LLM

Voice availability adapts to the model family you select:

• Gemini Voices (when a Gemini model is active): Sportsman, Customer support, Sarah, Brooke, Katie, Zemo, ajith, duaila, azj, ajz, sjl, brit, Swissen
• OpenAI Realtime Voices (when an OpenAI Realtime model is active): Alloy, Echo, Shimmer, Ash, Ballad, Coral, Sage, Verse, Cedar, Marin

Switching to a different provider in the LLM picker immediately refreshes the voice dropdown with the compatible set.

Voice Switching

Real-Time Changes:

• Instant Application: Voice changes take effect immediately
• WebRTC Reconnection: Brief system reconnection when switching voices
• Context Preservation: Conversation continues without interruption
• Avatar Consistency: Voice selection applies to current avatar

LLM Model Selection

Available Models

Provider Families in the Picker:

• OpenAI: GPT 4.1 mini, GPT 4.1, GPT 5, GPT 5 nano, GPT 5 mini
• Gemini: Gemini 2.5 Flash Lite, Gemini 2.5 Pro, Gemini 2.5 Flash
• OpenAI Realtime: GPT Realtime, GPT‑4o Realtime, GPT Realtime Mini (formerly “GPT‑4o Mini Realtime”)
• Groq: GPT OSS 20B, GPT OSS 120B, Qwen3‑32B, Moonshotai Kimi K2
• Creative Stack: Google Nano Banana, Veo 3.1, Sora 2 (used automatically for image/video generation prompts)

LLM picker groups models by provider so you can match speed, quality, and capability needs

Model Selection Strategy

Pick the right model based on your priority:

• Realtime conversations → OpenAI Realtime family for lowest latency, with GPT Realtime Mini as the fastest option
• Balanced quality + speed → OpenAI GPT 4.1 mini or Gemini 2.5 Flash Lite
• Premium reasoning → GPT 5, GPT 5 mini, Gemini 2.5 Pro, or Groq’s GPT OSS 120B
• Specialized workloads → Qwen3‑32B or Moonshotai Kimi K2 when you need alternative model behavior

Conversation History

History Management

Session Access:

• Previous Conversations: All past chat and voice sessions
• Organized Display: Conversations listed chronologically
• Session Details: Date, time, and duration information
• Quick Access: Click to load previous conversation context

Conversation history management showing organized sessions with “New Conversation” option and chronological listing of past interactions

Loading Previous Conversations

Context Integration:

• Load History: Bring previous conversation into current session
• Context Awareness: Historical information informs current responses
• Seamless Continuation: Resume conversations from where they ended
• Memory Integration: Past conversations influence agent memory

History Organization

Conversation Records:

• Today Section: Recent conversations from current day
• Date Organization: Conversations grouped by date
• Agent Identification: Which agents were used in each session
• Search Functionality: Find specific conversations quickly

Options Integration

Cross-Mode Compatibility

Universal Settings:

• Chat and Voice: Options work in both interaction modes
• Instant Switching: Change settings without mode interruption
• Consistent Experience: Settings maintained across mode switches
• Real-Time Application: Changes take effect immediately

Avatar Coordination

Avatar-Options Synchronization:

• Voice Matching: Selected voice applies to current avatar
• Agent Knowledge: Knowledge agent selection affects avatar responses
• Character Consistency: Avatar personality maintained across option changes
• Visual Coordination: Avatar appearance coordinated with voice selection

Scene Integration

Scene-Aware Options:

• All Scenes: Options available in every Hub scene
• Enhanced Features: Some scenes provide additional interaction capabilities
• Context Awareness: Scene selection may influence optimal settings
• Immersive Integration: Options work with interactive scene elements

Advanced Options Features

Smart Recommendations

Intelligent Suggestions:

• Agent Matching: System may suggest optimal agents for topics
• Voice Pairing: Recommended voice-avatar combinations
• Model Selection: Performance suggestions based on conversation type
• History Relevance: Relevant past conversations highlighted

Performance Optimization

System Efficiency:

• Connection Quality: Options menu shows connection status
• Model Performance: Real-time performance feedback
• Voice Quality: Audio quality indicators and optimization
• Response Timing: Model selection affects response speed

Customization Options

Personalization Features:

• Default Settings: Set preferred agents, voices, and models
• Quick Access: Favorite combinations for rapid selection
• Usage Patterns: System learns your preferences over time
• Workflow Optimization: Streamlined options based on usage

Best Practices

Optimal Configuration

Effective Settings:

• Match Purpose: Choose agents and models appropriate for conversation goals
• Voice Selection: Select voices that complement avatar appearance and role
• Model Efficiency: Balance capability needs with response speed requirements
• Context Utilization: Load relevant conversation history for continued discussions

Workflow Optimization

Efficient Usage:

• Quick Switching: Master rapid agent and voice switching for varied conversations
• History Leverage: Use previous conversations to build context and continuity
• Model Selection: Understand when to use different AI models for optimal results
• Setting Persistence: Learn which settings work best for different scenarios

Performance Management

System Efficiency:

• Connection Monitoring: Watch for connection quality indicators
• Resource Management: Balance features with system performance
• Update Awareness: Keep settings current with platform improvements
• Troubleshooting: Understand how settings affect performance and quality

Troubleshooting Options

Common Issues

Settings Problems:

• Options Not Loading: Refresh page or check connection status
• Agent Switch Fails: Verify agent selection and try again
• Voice Not Changing: Allow time for WebRTC reconnection
• History Not Loading: Check network connection and refresh if needed

Performance Issues

System Optimization:

• Slow Response: Try switching to GPT Realtime Mini or Gemini 2.5 Flash Lite
• Connection Problems: Monitor status indicators and connection quality
• Audio Issues: Check voice selection and browser audio permissions
• Memory Problems: Clear browser cache if options become sluggish

Configuration Troubleshooting

Setting Conflicts:

• Agent Mismatch: Ensure selected agent matches conversation topic
• Voice Compatibility: Verify voice works with current avatar
• Model Limitations: Check if selected model supports desired features
• History Access: Confirm conversation history is properly saved

Ready to optimize your avatar interactions? Use the options menu to customize your perfect conversation experience with the right combination of agents, voices, and AI models!